Resistance of a Human Serum-Selected Human ... - Journal of Virology

Vol. 67, No. 9

JOURNAL OF VIROLOGY, Sept. 1993, p. 5216-5225

0022-538X/93/095216-10$02.00/0 Copyright © 1993, American Society for MicrobiologY

Resistance of a Human Serum-Selected Human Immunodeficiency Virus Type 1 Escape Mutant to Neutralization by CD4 Binding Site Monoclonal Antibodies Is Conferred by a Single Amino Acid Change in gpl20 J. A. MCKEATING,lt*, J. BENNET,2 S. ZOLLA-PAZNER,3 M. SCHUT1rEN,4 S. ASHELFORD,S A. LEIGH BROWN,5 AND P. BALFE2 Chester Beatty Laboratories, Institute of Cancer Research, London SW3 6JB, 1 Virology Department, University College London Medical School, London WIP 6DB, 2 and Centre for HIVResearch, University of Edinburgh, Edinburgh EH9 3JN,5 United Kingdom; Research Service, Veterans Affairs Medical Center, New York, New York 10016, and Department of Pathology, New York University and Medical School, New York, New York 100103; and Laboratory of Immunology, National Institute of Public Health and Environmental Protection, 3720 BA Bilthoven, The Netherlands4 Received 1 March 1993/Accepted 2 June 1993

We have selected an HXB2 variant which can replicate in the presence of a neutralizing human serum. Sequencing of the gpl20 region of the env gene from the variant and parental viruses identified a single amino acid substitution in the third conserved region of gpl20 at residue 375 (AGT-*AAT, Ser--Asn; designated 375 S/N). The escape mutant was found to be resistant to neutralization by soluble CD4 (sCD4) and four monoclonal antibodies (MAbs), 39.13g, 1.5e, G13, and 448, binding to epitopes overlapping that of the CD4 binding site (CD4 b.s.). Introduction of the 375 S/N mutation into HXB2 by site-directed mutagenesis confirmed that this mutation is responsible for the neutralization-resistant phenotype. Both sCD4 and three of the CD4 b.s. MAbs (39.13g, 1.5e, and G13) demonstrated reduced binding to the native 375 S/N mutant gpl20. The ability to select for an escape variant resistant to multiple independent CD4 b.s. MAbs by a human serum confirms the reports that antibodies to the discontinuous CD4 b.s. are a major component of the group-specific neutralizing activity in human sera.

antibodies may play an important role in the cross-neutralizing response (21). The selection and characterization of neutralization-resistant escape mutants is one strategy for the identification of amino acid residues critical for the binding of neutralizing antibodies. We have selected escape variants which are resistant to both V3 MAbs and sCD4 (13, 15). To define the epitopes responsible for inducing cross-neutralizing antibodies in human sera further, we selected an escape variant that could replicate in the presence of a neutralizing human serum. Characterization of the neutralization sensitivity of the escape variant with a panel of MAbs mapping to the second variable loop (V2), V3, and the CD4 b.s. demonstrated that the variant was resistant to a number of MAbs mapping to the CD4 b.s., suggesting that antibodies specific for the CD4 b.s. in human sera are capable of crossneutralizing a nonhomologous HIV-1 isolate (HXB2) and may be sufficient to select for neutralization-resistant escape

Neutralizing antibody appears to be an important component of the protective immune response against human immunodeficiency virus type 1 (HIV-1) infection (1, 4-6). Sera from many HIV-1-infected individuals are able to neutralize a broad spectrum of virus isolates, indicating the presence of conserved neutralization epitopes (28, 37). The induction of antibodies with broad neutralizing capacity is a

primary goal of vaccine development strategies. Several reports have suggested that HIV-1 neutralizing activity is associated primarily with two regions of the gp120 envelope glycoprotein: the third hypervariable loop (V3) (8, 10, 29) and the CD4 binding site (CD4 b.s.) (11, 34). Generally, antibodies specific for V3 arise early after HIV infection and exhibit the ability to neutralize a limited number of HIV-1 isolates. However, a limited number of V3 monoclonal antibodies (MAbs) recognize conserved features of the loop and subsequently exhibit a broader neutralization profile (7, 10, 24). Later in the course of viral infection, antibodies capable of neutralizing a wider range of HIV-1 isolates are detectable (20, 28, 37). The appearance of such crossneutralizing activity coincides with the detection of antibodies capable of blocking gp120-soluble CD4 (sCD4) interaction (9, 11, 18, 34). A number of neutralizing MAbs derived both from infected individuals (9, 12, 26, 30, 36) and from recombinant gp120-immunized rats (3) which block gp120sCD4 binding have been identified, suggesting that such *

mutants.

MATERIALS AND METHODS Virus propagation and neutralization assays. (i) Transfection of H9 cells and viral stock production. H9 cells were washed twice in serum-free medium and resuspended at 106/ml in RPMI. Ten micrograms of HXB2 plasmid was transfected into 2 x 106 H9 cells by the use of Lipofectin as instructed by the manufacturer (GIBCO BRL, Gaithersburg, Md.). The transfected H9 cultures were monitored for viral infection by the detection of p24 antigen (as described

Corresponding author.

t Present address: Microbiology Department, University of Reading, Whiteknights, P.O. Box 228, Reading RG6 2AJ, United Kingdom. 5216

VOL. 67, 1993

RESISTANCE OF AN HIV-1 ESCAPE MUTANT TO NEUTRALIZATION

below) and by the appearance of multinucleated cells. Once the cells were producing virus (days 6 to 7 posttransfection), stocks of extracellular virus were made by mixing the infected cells with uninfected H9 cells in a 1:4 ratio. Fortyeight hours after the cocultivation, the supernatant was collected, aliquoted, and stored under liquid N2. Infectivity determinations were made by performing 10-fold serial dilutions of virus (50 pI) and incubation with 100 pI of C8166 cells at 2 x 105 cells per ml in microtiter plates at 37°C for 7 days. Plates were scored for the presence of p24 antigen, and the 50% tissue culture infective dose (TCID50) values were determined by the Karber formula. (ii) Neutralization assays. HIV (103 TCID50) in a volume of 40 ptl was incubated with 10 RI of a dilution of antibody, serum, or sCD4 under test at 37°C for 1 h. The virusantibody or virus-sCD4 mixture was then incubated with 100 pI of C8166 cells at a concentration of 2 x 105 cells per ml per well in a microtiter plate in triplicate. Five days postinfection, the wells were scored for the presence of syncytia, and the extracellular supernatant was collected from the wells by centrifugation, inactivated with 1% Empigen, and assayed for soluble p24 antigen as described previously (17). The lowest concentration of antibody or sCD4 resulting either in a complete blocking of syncytium formation or in a >90% reduction in p24 antigen production was defined as the reciprocal neutralization titer. (iii) Comparative neutralization of variant and wt virus by human sera. Human serum neutralization titers were obtained for both variant and wild-type (wt) viruses. The differences in the titers were compared by using a two-tail nonparametric test (Wilcoxon's rank sum) (see Fig. 7). MAbs and human sera. The following antibodies were used for neutralization studies and/or envelope binding studies: rat MAbs 10/54, 10/36e, and 11/85b, specific for amino acids 311 to 321 of the HXB10 V3 loop (14); rat MAb 41.1, recognizing a conformation-dependent epitope within the V3 loop (14); rat MAb 10/76b, mapping to a linear epitope in the V2 loop (17a); 38.1a, specific for amino acids 425 to 441, involved in CD4 binding (3); 39.13g, a rat MAb specific for an epitope overlapping the CD4 b.s. (3, 19); and human MAbs 588, 559, 448, 728, 654, 729 (12, 19), 1.5e, 2.1h (9, 35), and G13 (30), binding to closely related epitopes overlapping the CD4 b.s.; and murine MAb L120, specific for CD4 (Medical Research Council [MRC] AIDS Directed Programme [ADP] Repository). Sera were collected from 21 healthy CDC stage II and III seropositive males attending a sexually transmitted disease clinic as part of a longitudinal study of the natural history of HIV-1 infection (2). These individuals had not received antiretroviral therapy at the time of sampling. Antibody and sCD4 binding to detergent-solubilized viral gp 120. Virus-containing supernatants were inactivated with 1% Nonidet P-40 (NP-40), which does not irreversibly denature gpl20/160 (23). The concentrations of gpl20 present in the preparations were determined by a twin-site enzyme-linked immunosorbent assay (ELISA) as previously described (22, 23), using recombinant gpl20 (rgpl2O; BH10 clone), expressed in and purified from CHO cells, as a reference standard. Briefly, detergent-solubilized gpl20 from wt, variant, and 375 S/N (see Results for definition) site-directed mutant viral stocks was allowed to bind to the solid phase via the capture antibody D7324 (anti-C terminus; Aalto-Bioreagents, Dublin, Ireland) at an input concentration of 50 ng/ml. The abilities of MAbs and sCD4 to bind to the captured gpl20 were assessed by using methods previously published (13, 19). In each case, the result is expressed as a

5217

ratio of the optical density of the mutant protein to that of the wt protein and is termed the relative binding index. MAb cross-competition analysis for rgpl2O binding. The human CD4 b.s. MAbs and three control rat MAbs binding to the V2, V3, and fourth conserved (C4) regions of gpl20 were tested for the ability to compete with iodinated preparations of MAb 39.13g for binding to D7324-bound rgpl20. The unlabeled MAbs, at a concentration sufficient to saturate the gpl20 (5.0 ,ug/ml), were mixed with an equal volume of 39.13g, at a concentration resulting in half-maximal binding (0.16 ,ug/ml), and the mixture was incubated with gpl20 for 1 h. The amount of 39.13g bound to gpl20 in the presence or absence of the test MAbs was determined. Cell surface MAb and sCD4 binding. Infected H9 cells were washed twice in complete phosphate-buffered saline (PBS)-1% fetal calf serum-0.05% sodium azide (WB), resuspended at a concentration of 107 cells per ml in WB, and chilled on ice. MAbs or sCD4 were diluted in WB, and 100 ,ul was mixed with 100 ,u of cells (106 cells) and incubated in a V-bottom microtiter plate for 30 min. Cells were washed with WB by centrifugation, and the bound MAb was detected with a 1/40 dilution of fluorescein isothiocyanateconjugated anti-rat immunoglobulin G (IgG) or anti-human IgG (SeraLabs; Crawley, Sussex, United Kingdom) in WB. Bound sCD4 was detected with CD4-specific MAb L120 (MRC ADP Repository) and fluorescein isothiocyanate-conjugated anti-mouse IgG. After a 30-min incubation on ice with the conjugate, the cells were washed three times with WB and inactivated by resuspension in 500 RIu of 1% paraformaldehyde in PBS at 4°C overnight. Cells were analyzed on a FACStar, using Lysis software (Becton Dickinson). The results shown are the mean fluorescence intensities obtained for one experiment. However, all experiments were repeated three times. Cassette vector construction and site-directed mutagenesis. The HXB2 insert of plasmid pHXB2gpt was transferred into a vector lacking XbaI sites by ligation into the NheI site of a pSPTBM20 vector (Boehringer Mannheim), from which the polylinker had previously been excised by double digestion with SmaI and EcoRV and self-ligation. A derivative, MCS, was constructed by introducing a silent site-directed mutation in the gpl60 gene to generate an XbaI site and two nonsilent changes in the nonfunctional vpu gene to generate a NotI site. Virus recovered upon transfection of the modified plasmid was able to infect H9 cells with the same infection kinetics as the parental HXB2. Infection of 106 H9 cells with viral doses of HXB2 and HXB2-MCS containing 200 pg of virion p24 resulted in soluble p24 production at 4, 6, 8, and 10 days postinfection of 7.6, 120.0, 390.0, and 510.0 ng/ml and 5.0, 185.0, 425.0, and 525.0 ng/ml, respectively. A deleted form of MCS made by double digestion of the construct with NheI and MamI followed by Klenow fill-in of the NheI end and self-ligation. This digestion excised a 295-bp piece of the envelope gene including the V4, C4, and V5 domains (see Fig. 3), and the plasmid was shown to be noninfectious (data not shown). Generation and cloning of wild-type and variant gpl20 molecules. Single molecules of gpl20 were amplified by a limit dilution method (31) using nested primers specific for the vpu region (sense) and for the gp4l molecule (antisense). Initial amplification was performed in a volume of 20 ,u with primers TCA TCA AGT TTC TCT AC/TC AAA GC (5' sense, outer; positions 5568 to 5590 (23a) and TCC CAC TCC ATC CAG GTC (3' antisense, outer; positions 7664 to 7647) (23a), using 30 cycles of polymerase chain reaction (PCR) amplification (94°C, 40 s; 50°C, 35 s; 72°C, 210 s) with

5218

McKEATING ET AL.

J. VIROL.

TABLE 1. Neutralization sensitivities of wt and variant viruses Neutralization titer Ligand

Epitope

(,ug/ml or reciprocal dilutiona) wt

41.1 10/54 10/36e

11/85b 10/76b 39.13g sCD4 QC5 (selecting serum)

V3 V3 V3 V3 V2 CD4 b.s.

0.4 + 0.05 9.4 ± 1.8 1.2 + 0.4 1.3 + 0.5 0.9 ± 0.2 5.4 ± 0.8 2.5 ± 0.4 1/640

Variant

0.5 8.4 0.8 1.1 1.2

± 0.06 ± 1.6 ± 0.3

RESULTS Selection procedure. Experiments were designed to monitor the kinetics of the appearance of HIV-1 escape mutants propagated in the presence of a HIV-1-positive human serum. The serum chosen, QC5 (16), neutralized HXB2 at a serum dilution of 1/640 and contained antibodies capable of blocking the gpl2O-sCD4 interaction (data not shown). Virus

± 0.3 ± 0.3 A

NRb NR 1/150

a Reciprocal dilution is given for QC5. b NR, 90% reduction in p24 antigen production was not achieved for the variant virus with either 39.13g or sCD4. At 20 ,ug/ml, the variant was neutralized by 10 and 53% by 39.13g and sCD4, respectively.

-7_

.(A) 0.6 0.50.40.7

r.

- ain

0.3-

0

0.2-

16 ng of each primer, 200 ,uM each deoxynucleoside triphosphate (dNTP), 50 ,uM TMAC, 1 x PFU of polymerase buffer 3, and 0.5 U of polymerase (Stratagene). To prevent degradation, the first-round product was frozen immediately upon completion of the reaction. One microliter of this first-round product was transferred to a second tube containing a volume of 100 ,ul with primers GTA GCA TTA GCG GCC GCA ATA ATA ATA GCA ATA G (5' sense, NotI inner; positions 5635 to 5668) (23a) and GTT CTA GAG ATT TAT TAC TCC (3' antisense, XbaI inner; positions 7631 to 7611) (23a), using 25 cycles of amplification (94°C, 35 s; 55°C, 25 s; 72°C, 150 s) with 80 ng of each primer, 200 ,uM each dNTP, 50 ,uM TMAC, lx PFU polymerase buffer 3, and 2.5 U of PFU polymerase. Ten microliters of the PCR product was visualized on an agarose gel, and the remainder was purified by a Gene-Clean (Bio 101, La Jolla Calif.) procedure. After digestion with NotI and XbaI (Promega Biotec), and PCR product was ligated into NotI-XbaI-digested pHXB2MCSAenv and cloned into Escherichia coli SURE (Stratagene). Colonies were screened for full-length clones by PCR and full-length clones were sequenced to confirm the insert genotype. Site-directed mutation and cloning of gp120 molecules. Site-directed mutagenesis of the HXB2 envelope was performed by a modification of the method of Wolfs et al. (38), using PFU polymerase in place of Taq polymerase. Two 20-pl PCR reactions were set up as described above, with 25 ng of pHXB2gpt as the template, using either the 3' antisense XbaI primer and a sense primer encoding the desired 375 mutation or the 5' sense NotI inner primer and a second, antisense, 375 primer. The use of this relatively large amount of template DNA allowed us to reduce the number of cycles performed in the PCR and hence the frequency of polymerase-introduced error; the PCR was run for 10 cycles, using the temperature profile described above for the second round. The PCR products were gel purified by a Gene-Clean procedure, and the two products were pooled into a 100-,ul PCR mix containing the 5' sense Notl inner primer and the 3' antisense XbaI primer; this mixture was reamplified for a further 10 cycles, using the same temperature profile as before. The final, full-length product was separated from the two shorter products on a 1.5% agarose gel, purified, cloned, and screened as described above. The presence of the desired change was verified by sequencing.

m 0.1

~~~~~-0---